Prior beliefs for predicting movements

Timing and motor control seem effortless in daily life, yet lab measurements of movement timing often show biases and variability. We resolve this conundrum through probability theory and provide a biophysical mechanism for how single neurons encode prior probabilities of motor timing to drive precise behaviors. We follow these single neuron dynamics to the microcircuit and population levels, where nonlinear dynamics shape neural trajectories into topological manifolds, which provide substrates for prior knowledge in complex motor timing behaviors. At the end we unify these different levels of analysis to understand how the brain optimally controls the timescales of our movements. 

Chair: Daan Brinks